Kubernetes architecture: control plane and worker nodes explained

Two Kinds of Machines in a Cluster

A Kubernetes cluster has two roles: the control plane (the brain) and worker nodes (the muscle).

Control Plane Components

API Server — every request (kubectl, other components) goes through this REST gateway. It is the single entry point to the cluster state.

etcd — a distributed key-value store. This is where all cluster state lives. Back this up. Losing etcd means losing your cluster.

Scheduler — watches for unscheduled Pods and assigns them to a suitable node based on resource availability and constraints.

Controller Manager — runs reconciliation loops. The ReplicaSet controller, for example, notices if a Pod dies and creates a replacement.

Worker Node Components

kubelet — an agent on every node. It receives Pod specs from the API server and tells the container runtime to start or stop containers.

kube-proxy — manages network rules on the node so Pods can communicate across the cluster.

Container Runtime — the engine that actually runs containers (containerd is the default; Docker is no longer directly supported).

# Check your cluster nodes and their roles
kubectl get nodes

# Describe a node to see its components and capacity
kubectl describe node <node-name>